Windows desktop AI visual effects engine that generates filters, particles, models and scenes from natural language, with parameters adjustable in real time and interaction with cameras and layers.
A Bilibili creator "_OXOX_" (formerly Oxoxxidane, about 7K followers, making AE/motion content) spent nearly half a year building an AI visual effects engine alone. On 2026-09-10 he posted the launch video on his own Bilibili channel, "Generate any AE effect plugin in 8 minutes with one click!" (25K views and 1,968 favorites, far beyond his follower count), offering 1,200 free credits to get people to download. The product's pitch is "it doesn't make decisions for you, all parameters adjustable in real time", and it comes with a community effects gallery (user-generated effects can be made public and commented on); in 2026-10 he also released the Code2AE agent skill on GitHub for tools like Claude/Codex. The cold start relied entirely on the launch on Bilibili and the favoriting and sharing within the AE creator circle.One-line growth story, compiled by us from public sources; speculation is marked separately below.
The main engine is the author's own launch video on Bilibili: the After Effects motion creator _OXOX_ posted the launch video, which got 25,000 views and 1,968 saves, far above his 7,000 followers; the Similarweb public page has no channel data for this site.
The lubricant is the credit system: new users get 1,200 credits, and when they run out they must come back to the site to buy a credit pack from $2.90 or a subscription from $9.90/month; the client and the AE plugin also require logging in on the site and creating an API Key.
By our estimate, most visits are users who installed the client coming back to log in and buy credits; visit counts are estimated, and about 15 of every 100 visits click through to the Creem checkout page, which is exactly the ratio assumed when estimating total visits.
Model: Lenny's racing growth framework. Engine = the author's own launch on Bilibili (an AE motion creator posts a launch video that gets saved and shared in the AE creator community); lubricant = the credit system and the desktop client (1,200 free credits, and when they run out users return to the site to buy a credit pack or subscription, and the client requires logging in and creating an API Key).
Real data: The Similarweb public page has no global rank or channel data for this site, and no country, device or bounce rate either; 5,312 visits (about 5,300) to the Creem checkout page come from quietmoney checkout page data.
Supporting evidence: The author _OXOX_ has about 7,000 followers on Bilibili, and his Bilibili description says this is a personal project he has developed continuously for nearly half a year; the launch video has 25,463 views, 1,816 likes, 1,968 saves, 287 shares and 140 comments (Bilibili API); the domain was registered in 2026; pricing is credit packs at $2.90 (1,500) and $12 (7,500), and subscriptions Starter $9.90/month, Pro $19.90/month and Max $39.90/month; the community gallery has user-published works and comments; the author later also published on GitHub the Code2AE skill that depends on the ArbiFX plugin, with 18 stars, which is not counted in last month's visits.
Inference: The 35,100 total visits have no global rank to use and are estimated as checkout page visits of 5,300 divided by a typical small-site ratio of 15%, and may be off by several times, so the channel circle carries a ≈; for the same reason the 15% for 'site visits to checkout page' is an assumption, not a measured figure. The channels have no data, so only one circle is drawn. How each route splits is an inference: 35,100 = client users coming back to log in and buy credits 30,000 + seeing the Bilibili launch video 3,600 + seeing effects shared by others 1,500; the launch video has only 25,000 views in total, so this route is given a small share. The profile once recorded referrals at 95.5%, an older basis, used only to judge that the bulk is probably client login redirects and links in the video description. That 15% visit after watching the video and 10% after seeing a work are guesses.
Checkout: creem.io
New-user source: Other (Direct / referral / email / paid). Referrals 95.5%, the rest unknown; there are no search terms, referral site breakdown or country data, so the sources cannot be split further.
Top channel measured by Similarweb, others estimated by rank
| Attributable new users, last month | Share |
|---|---|
| Other | 100% |
Direct + brand search 0%. Direct and brand search count as viral only when the site shows sharing mechanics; the largest share wins, under 5% means Other.
Mainly Chinese, plus Japanese and English
About 5,312 last month (Similarweb estimate), #526 of 606; the month before was 0. Opening checkout is not paying, so this is not revenue.
New-user source: Other (Direct / referral / email / paid). Mostly referrals, email or paid, or mostly direct with no clear source Referrals 95.5%, the rest unknown; there are no search terms, referral site breakdown or country data, so the sources cannot be split further.
Credit packs $2.90 (1,500) and $12 (7,500); subscriptions Starter $9.90/month, Pro $19.90/month, Max $39.90/month
Dossier confidence: low. Parts 1 and 2 come from reading the site and Similarweb data; part 3 from public reporting, founder posts, Wayback snapshots, WHOIS and similar, with speculation marked.
The number: checkout visits is Similarweb's estimate of visits from the site to a checkout payment page; opening checkout is not paying, so this is not revenue. Small sites land on fixed buckets.
Data through last month. Page updated 2026-10-31.